2102
J. Phys. Chem. B 2002, 106, 2102-2108
Evidence for Anisotropic Coupling between the Protein Environment and the Copper Site in Azurin from Resonance Raman Spectroscopy M. Adam Webb† and Glen R. Loppnow* Department of Chemistry, UniVersity of Alberta, Edmonton, Alberta T6G 2G2, Canada ReceiVed: October 1, 2001; In Final Form: NoVember 29, 2001
A key problem in biological electron transfer is the independent measurement of the parameters upon which the rate constant for electron transfer depends. In this paper, we report evidence that quantitative resonance Raman spectroscopy may be such an independent probe of the electronic coupling parameter in proteins. The resonance Raman cross-sections of four species of azurin have been measured throughout the ca. 600-nm absorption band. The total resonance Raman intensity, as quantified by the cross-sections, is very similar for the four azurins in the 200-700 cm-1 region. However, the resonance Raman intensity is distributed differentially among the vibrational modes, sometimes also accompanied by small frequency shifts. Correlations are attempted between this re-distribution of resonance Raman intensity and Cu-S bond length, Cu out-ofplane distance, backbone structure, and protein environment. Of these four parameters, the only significant correlation exists with the protein environment. A detailed analysis of the residues within 10 Å of the copper site demonstrates a strong, anisotropic coupling between the copper site and Trp48, a residue previously implicated in intramolecular electron transfer pathways of azurin.
Introduction Blue copper proteins, such as azurin, are involved in respiratory and photosynthetic electron transport chains1-3 and have been used extensively as models for long-range protein electron transfer.4-16 The active site of azurin is composed of a Cu2+ ion strongly coordinated to two histidines and one cysteine, and weakly coordinated to a methionine and the carboxyl oxygen of a glycine.3 Azurin exhibits a strong absorption band ca. 600 nm that has been assigned as a charge transfer band due to its high oscillator strength, although recent evidence17,18 suggests the character is more ππ* and little charge is transferred. Time-resolved absorption spectroscopy of Rumodified5-11 or 1-thiouredopyrene-3,6,8-trisulfonic (TUPS) acid-modified15,16 azurin has shown a dependence of electron transfer rates on protein composition and structure. This dependence has been modeled theoretically as a sequence of through-space, through-bond, and through-hydrogen bond couplings.5-7 Similar studies on azurin have examined particular electron transfer pathways by combining pulse radiolysis experiments with site-directed mutagenesis of residues between the electron donor and acceptor.12,13 All of these studies have shown that the molecular nature of the medium is critical for predicting rates of electron transfer. To understand the molecular mechanism of electron transfer, the roles of distance, medium, orientation, and protein structure must be clearly elucidated. The electron transfer rate constant is dependent on the electronic coupling, HDA, between electron donor and acceptor at the transition state, given by the Marcus equation19
k)
2 2 2π HDA e-(∆G°+λ) /4λRT p (4πλRT)1/2
(1)
where ∆G° is the standard free energy of reaction and λ is the * To whom correspondence should be addressed. E-mail:
[email protected]. † Current address: Department of Chemistry, University of Idaho, Moscow, ID 83844-2343.
nuclear reorganization energy. For large separation of donor and acceptor, such as in long-range electron transfer in DNA and proteins, HDA is expected to decay exponentially
HDA ) H0DAe-β(r-r0)
(2)
where H0DA is the coupling at r0, r0 is the van der Waals separation between electron donor and acceptor, r is the actual distance between electron acceptor and donor, and β is an empirical decay constant. A number of both experimental and theoretical studies have suggested that the value of β depends on the particular pathway of the electron transfer5-7,19 through the intervening medium in large molecules. However, the experimental studies supporting this approach have relied primarily on measurements of the rate constant as a function of distance between donor and acceptor. A recent study20 has indicated that the parameters upon which the rate constant for electron transfer depends may not be independent; i.e., changing the distance between donor and acceptor may not only affect the coupling but may also affect the reorganization energy and the driving force of the electron transfer reaction. Resonance Raman spectroscopy may provide information about the electron transfer pathways in proteins and other biological molecules. In resonance Raman spectroscopy, vibrational Raman scattering is excited with radiation at an energy coincident with an electronic transition. The resulting resonance Raman frequencies reflect the normal mode character, while the intensities reflect excited-state structure and dynamics. In proteins, chromophores can be selectively probed by exciting within an absorption band of the chromophore far from any protein absorption. Previous work in our group has extended this idea by varying the protein environment and examining the resulting resonance Raman spectrum of the chromophoric active site, in effect to use the chromophore to project out the important couplings between the active site and protein environment.21-24 In this paper, the absorption spectra and quantitative resonance Raman excitation profiles of four azurins from Pseudomonas
10.1021/jp013665b CCC: $22.00 © 2002 American Chemical Society Published on Web 02/05/2002
Anisotropic Coupling in Azurins
J. Phys. Chem. B, Vol. 106, No. 8, 2002 2103
aeruginosa (PA), Alcaligenes denitrificans (AD), and Alcaligenes xylosoxidans (AX I and AX II) have been analyzed. The good correlation of spectral differences with protein composition indicates that resonance Raman spectroscopy is sensitive to the coupling of the copper to the protein environment. Surprisingly, the sensitivity of the resonance Raman spectra appears to be directionally anisotropic; i.e., the spectrum is more sensitive to amino acid changes in some parts of the protein environment than in others. This result suggests an anisotropic long-range coupling between the copper active site and the protein, and that resonance Raman spectroscopy may be sensitive to hardwired electron transfer pathways important for this protein’s function. Methods Experimental Details. The two azurins (AX I and AX II) from A. xylosoxidans (NCIMB 11015) were isolated and purified from literature procedures25-27 with slight modifications24 in the cell rupture and column chromatography steps. Column chromatography was performed using Whatman CM-52 and Sephadex G-50 columns until the purity ratio (A280/A620) was 2.0-4.9 for AX I and 3.1-4.7 for AX II. Typical yields were 4-24 mg of AX I and 4-8 mg of AX II per 100 g of cell paste. Samples were prepared for resonance Raman experiments by quantitative dilution with a cacodylate buffer solution (0.51.0 M cacodylate, 0.01 M TRIS-HCl, pH 8.7). Addition of cacodylate buffer did not have a noticeable effect on the absorption or resonance Raman spectra of azurin. Roomtemperature resonance Raman spectra were taken as described previously21-24,28,29 using 250-300 µL aqueous samples (0.01 M Tris-HCl, pH 8.7, 0.45-0.83 M and 0.34-0.75 M cacodylate for AX I and AX II, respectively) having an absorbance of 1-11 OD/cm and 2-5 OD/cm ca. 620 nm for AX I and AX II, respectively. Bleaching of the sample was accounted for by measuring the absorption at 560 nm (560 ) 2720 M-1 cm-1 for AX I, 560 ) 2710 M-1 cm-1 for AX II). For all samples, the overall bleaching was
